Signed Neuron with Memory: Towards Simple, Accurate and High-Efficient ANN-SNN Conversion

Authors: Yuchen Wang, Malu Zhang, Yi Chen, Hong Qu

IJCAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct experiments on challenging datasets including CIFAR10 (95.44% top-1), CIFAR100 (78.3% top-1) and Image Net (73.16% top-1). Experimental results demonstrate that the proposed method outperforms the state-of-the-art works in terms of accuracy and inference time.
Researcher Affiliation Academia Yuchen Wang , Malu Zhang , Yi Chen and Hong Qu School of Computer Science and Engineering, University of Electronic Science and Technology of China yuchenwang@std.uestc.edu.cn, maluzhang@uestc.edu.cn, chenyi@std.uestc.edu.cn, hongqu@uestc.edu.cn
Pseudocode No The paper describes mathematical equations for the neuron model (Eq. 7-11) but does not contain structured pseudocode or algorithm blocks (clearly labeled algorithm sections or code-like formatted procedures).
Open Source Code Yes The code is available at https://github.com/ppppps/ ANN2SNNConversion SNM Neuron Norm.
Open Datasets Yes We use VGG and Res Net network structures to conduct experiments on CIFAR10, CIFAR1001, and Image Net20122 datasets. ... 1https://www.cs.toronto.edu/ kriz/cifar.html 2https://image-net.org/challenges/LSVRC/2012/
Dataset Splits No The paper mentions using a "training set" to obtain the maximum activation value for neuron-wise normalization, but it does not specify any dataset splits (e.g., percentages or counts) for training, validation, or testing.
Hardware Specification No The paper discusses energy estimation using theoretical values for multiplication and addition operations (e.g., 4.6 pJ and 0.9 pJ from [Horowitz, 2014]), but it does not specify any concrete hardware details such as exact GPU/CPU models, processor types, or memory used for running its experiments.
Software Dependencies No The paper mentions general algorithms and techniques like SGD and Kaiming normal initialization but does not provide specific software dependencies with version numbers (e.g., PyTorch 1.x, TensorFlow 2.x, Python 3.x).
Experiment Setup Yes The initialization of ANN network parameters adopts Kaiming normal initialization [He et al., 2015], and the network training adopts SGD algorithm with 0.9 momentum followed by milestones learning rate decay. The L2 penalty with a value of 5e 4 is also added.